A batting average in cricket is simply the runs scored divided by the number of innings played. Remember only the innings in which a batsman got out are considered. Thus if a batsman scored 3203 runs in the 100 innings he played and he was not-out 20 times, will result in an average of 3203/80 i.e. a shade over 40. Average is traditionally an important statistic to measure a player’s performance or effectiveness. It can often be the reason for a player to be picked or dropped from the squad.
Most of us remember Michael Bevan as the guy who maintained an average of more than 50 runs for a long time. However, what I remember is that he used to take minimal risks in the fag end of the innings to protect his wicket and thus increase his batting average! Indian skipper Dhoni also averages over 50. Well, if these guys have such good averages then what is the problem?
The Problem with averages as a parameter to judge a batsman’s performance
The problem is that the Not Outs can inflate the averages. They do not show the average contribution per game (i.e. the average runs scored per innings). Averages are so much in use to determine/gauge a player’s performance; hence averages if inflated by the number of not outs can present an incorrect picture. Let’s have a look at some data to test this out:
Michael Bevan played 232 ODI’s, 196 innings and scored 6912 runs and was not out 67 times. Thus, his batting average is a staggering 53.58 (6912/129); however his contribution per innings is 6912/196 = 35.26. Compare 53.58 to 35.26, a percent change of 34%, huge isn’t it?
In case of MS Dhoni, he has played 211 ODI’s, 188 innings and scored 6908 runs, he was not out 53 times. His batting average is 51.17, however runs scored per innings is 36.74. Again a percent change of over 28%. Now does it matter to the team if a player scores 51 or 51* (not out). All that matters is the overall contribution to the team i.e. the runs scored by the batsman in a given match.
If we look at Ganguly’s record, his batting average is 41.02 and the average runs scored per innings is 37.87. Thus if you compare with Dhoni, though Ganguly’s average is far less but the average contribution is still more. Same in the case of Gambhir whose average is 40.94 but average runs scored per match is 37.6 which is also more than Dhoni’s. Any bells ringing? There are many more examples.
Agreed the batting position also affects the batting averages, top order batsmen usually get more chances to score as compared to middle order and lower order batsmen. But the point I am trying to raise here is that batting averages are not always the correct indicator of a player’s performance.
What went into the analysis?
I took a random sample of 20 cricketers from India, Pakistan, Sri Lanka and Australia and carried out a correlation between a player’s average and the percentage of innings in which a player was not out.
A correlation is the relationship or connection between two or more things.
For example Dhoni’s career average is 51.17 and he played 211 ODI’s and 188 innings with 53 not outs, thus the %age of not outs is (53/188)*100 i.e. he was not out in 28.19% of his innings.
The correlation graph looks like this:
The graph shows that with increasing %age of not-outs the batting average also increases. However the correlation is not that strong (R²=0.3178) but still it is a positive correlation. The correlation cannot be that strong because ultimately the averages boil down to a player’s batting performances.
Probably it is high time that cricket management and statisticians start paying more attention to the average runs scored per match and not just rely on the averages!